Top Banner
CLARKIN ET AL. BORDERLINE PERSONALITY DISORDER THE PERSONALITY DISORDERS INSTITUTE/BORDERLINE PERSONALITY DISORDER RESEARCH FOUNDATION RANDOMIZED CONTROL TRIAL FOR BORDERLINE PERSONALITY DISORDER: RATIONALE, METHODS, AND PATIENT CHARACTERISTICS John F. Clarkin, PhD, Kenneth N. Levy, PhD, Mark F. Lenzenweger, PhD, and Otto F. Kernberg, MD The Personality Disorder Institute/Borderline Personality Disorder Re- search Foundation randomized control trial (PDI/BPDRF RCT) is a con- trolled outcome study for borderline personality disorder (BPD), in which 90 participants were randomized to one of three manualized and moni- tored, active psychosocial treatment conditions. These treatments are: (a) Transference-Focused Psychotherapy (TFP; Clarkin, Yeomans, & Kernberg, 1999), a treatment for BPD based on object-relational and psy- choanalytic principles first applied to BPD by Kernberg (1996), notable for its particular emphasis on interpretation of object relations activated in the ongoing therapeutic relationship; (b) Dialectical Behavior Therapy (DBT; Linehan, 1993), a popular treatment for BPD, with evidence of effi- cacy (Linehan, Armstrong, Suarez, Allmon, & Heard, 1991) that empha- sizes a balance between acceptance and change in its combination of cognitive-behavioral and Zen principles; and (c) supportive psychother- apy (Rockland, 1992), another object-relational and psychoanalytically based treatment for BPD which, in contrast to TFP, eschews transference interpretation and places primary emphasis on development of a collabo- 52 Journal of Personality Disorders, 18(1), 52-72, 2004 © 2004 The Guilford Press From the Personality Disorders Institute and Department of Psychiatry, Weill Medical College of Cornell University (J.F.C., O.F.C.), City University of New York (K.N.L.), and State University of New York at Binghamton (M.L.). This research was supported, in part, by grants from the Borderline Personality Disorder Re- search Foundation (O.F.K, J.F.C.). The authors thank Jack Barchas, MD for institutional support and acknowledge the technical assistance of Kenneth L. Crutchfield, PhD, Jill C. Delaney, MSW, Simone Hoermann, PhD, Joel McClough, PhD, and Maya Kirschner, PhD for their help in conducting assessments, and James Hull, PhD, for his help in organizing and maintaining the data. They also acknowledge the consultation of Marsha Linehan, PhD, to this study and Heidi Heard, PhD, in training with some of the assessment instruments. They also thank their colleagues for their contributions as therapists in the study; Drs. Ann Appelbaum, Barbara Stanley, and Frank Yeomans for pro- viding supervision as treatment cell team leaders, members of the Personality Disorders Insti- tute, and the patients for their participation in the project. Address correspondence to Dr. John F. Clarkin, Personality Disorders Institute, Macy Villa, The New York Presbyterian Hospital-Weill Medical College of Cornell University, Westchester Division, 21 Bloomingdale Rd., White Plains, NY 10605; Email: [email protected].
21

CLARKIN ET AL.BORDERLINE PERSONALITY DISORDER · the personality disorders institute/borderline personality disorder research foundation randomized control trial for borderline personality

Sep 20, 2020

Download

Documents

dariahiddleston
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Page 1: CLARKIN ET AL.BORDERLINE PERSONALITY DISORDER · the personality disorders institute/borderline personality disorder research foundation randomized control trial for borderline personality

CLARKIN ET AL.BORDERLINE PERSONALITY DISORDER

THE PERSONALITY DISORDERSINSTITUTE/BORDERLINE PERSONALITYDISORDER RESEARCH FOUNDATIONRANDOMIZED CONTROL TRIAL FORBORDERLINE PERSONALITY DISORDER:RATIONALE, METHODS, AND PATIENTCHARACTERISTICS

John F. Clarkin, PhD, Kenneth N. Levy, PhD,Mark F. Lenzenweger, PhD, and Otto F. Kernberg, MD

The Personality Disorder Institute/Borderline Personality Disorder Re-search Foundation randomized control trial (PDI/BPDRF RCT) is a con-trolled outcome study for borderline personality disorder (BPD), in which90 participants were randomized to one of three manualized and moni-tored, active psychosocial treatment conditions. These treatments are:(a) Transference-Focused Psychotherapy (TFP; Clarkin, Yeomans, &Kernberg, 1999), a treatment for BPD based on object-relational and psy-choanalytic principles first applied to BPD by Kernberg (1996), notablefor its particular emphasis on interpretation of object relations activatedin the ongoing therapeutic relationship; (b) Dialectical Behavior Therapy(DBT; Linehan, 1993), a popular treatment for BPD, with evidence of effi-cacy (Linehan, Armstrong, Suarez, Allmon, & Heard, 1991) that empha-sizes a balance between acceptance and change in its combination ofcognitive-behavioral and Zen principles; and (c) supportive psychother-apy (Rockland, 1992), another object-relational and psychoanalyticallybased treatment for BPD which, in contrast to TFP, eschews transferenceinterpretation and places primary emphasis on development of a collabo-

52

Journal of Personality Disorders, 18(1), 52-72, 2004© 2004 The Guilford Press

From the Personality Disorders Institute and Department of Psychiatry, Weill Medical Collegeof Cornell University (J.F.C., O.F.C.), City University of New York (K.N.L.), and State Universityof New York at Binghamton (M.L.).This research was supported, in part, by grants from the Borderline Personality Disorder Re-search Foundation (O.F.K, J.F.C.).The authors thank Jack Barchas, MD for institutional support and acknowledge the technicalassistance of Kenneth L. Crutchfield, PhD, Jill C. Delaney, MSW, Simone Hoermann, PhD, JoelMcClough, PhD, and Maya Kirschner, PhD for their help in conducting assessments, andJames Hull, PhD, for his help in organizing and maintaining the data. They also acknowledgethe consultation of Marsha Linehan, PhD, to this study and Heidi Heard, PhD, in training withsome of the assessment instruments. They also thank their colleagues for their contributionsas therapists in the study; Drs. Ann Appelbaum, Barbara Stanley, and Frank Yeomans for pro-viding supervision as treatment cell team leaders, members of the Personality Disorders Insti-tute, and the patients for their participation in the project.Address correspondence to Dr. John F. Clarkin, Personality Disorders Institute, Macy Villa,The New York Presbyterian Hospital-Weill Medical College of Cornell University, WestchesterDivision, 21 Bloomingdale Rd., White Plains, NY 10605; Email: [email protected].

Page 2: CLARKIN ET AL.BORDERLINE PERSONALITY DISORDER · the personality disorders institute/borderline personality disorder research foundation randomized control trial for borderline personality

rative engagement with the patient to foster identity development. Pa-tients received medication, if clearly indicated, according to thetreatment algorithm developed by Soloff (2000). This article describes thesignificance and rationale of the study and the overall design, methods,plan of analysis, and demographic characteristics of the recruited sampleof patients.

Borderline personality disorder (BPD) constitutes one of the most importantsources of long-term impairment in both treated and untreated populations(Widiger & Weissman, 1991). BPD is a prevalent, chronic, and debilitatingsyndrome associated with high rates of medical and psychiatric use of ser-vices (Lenzenweger, Loranger, Korfine, & Neft 1997; Torgersen, Kriglen, &Cramer, 2001; Skodol, Gunderson, Pfohl, Widiger, Livesley, & Siever, 2002).Approximately 11% of psychiatric outpatients and 19% of inpatients metthe Diagnostic and Statistical Manual (4th ed.) criteria for BPD (Kass,Skodol, Charles, Spitzer, & Williams, 1985), the majority of whom arewomen.

Suicidal and self-injurious behavior is particularly prevalent with BPDpatients, with rates ranging from 69% to 75% (McGlashan, 1986; Stone,1993; Cowdry, Pickar, & Davies, 1985; Clarkin, Widiger, Frances, Hurt, &Gilmore, 1983). BPD is substantially comorbid with other personalitydisorders (Nurnberg et al., 1991; Zimmerman & Coryell, 1990) and withAxis I disorders (Fyer, Frances, Sullivan, Hurt, & Clarkin, 1988). BPDnegatively affects the treatment efficacy for a number of Axis I disorders(see Clarkin, 1996), and is less responsive to pharmacotherapy (Soloff,2000).

PREVIOUS PSYCHOTHERAPY RESEARCH ON BPDPsychotherapy is the most widely practiced technique for treating border-line patients and a recent meta-analysis by Perry, Banon, and Ianni(1999) suggests that psychotherapy is an effective treatment for person-ality disorder and may be associated with up to a sevenfold faster rate ofrecovery in comparison with the natural history of disorders. Althoughpsychotherapy is the recommended primary technique for treating bor-derline patients (Oldham et al., 2001) and findings like Perry and col-leagues’ are encouraging, few studies have actually examined theeffectiveness of particular treatments for borderline patients (Bateman &Fonagy, 1999; Blum, Pfohl, St. John, Monahan, & Black, 2002; Clarkin,Foelsch, Levy, Hull, Delaney, & Kernberg, 2001; Cookson, Espie, & Yates,2001; Linehan et al., 1991; Linehan et al., 1999; Ryle & Golynkina, 2000;Stevenson & Meares, 1992). In our review we found only five publishedrandomized controlled trials (Bateman & Fonagy, 1999; Koons et al.,2001; Linehan et al., 1991; Linehan et al., 1999; Munroe-Blum &Marziali, 1995), with only two treatments—a psychodynamic day hospi-tal program and dialectical-behavioral therapy (DBT)—having shown

BORDERLINE PERSONALITY DISORDER 53

Page 3: CLARKIN ET AL.BORDERLINE PERSONALITY DISORDER · the personality disorders institute/borderline personality disorder research foundation randomized control trial for borderline personality

acute efficacy for treating BPD (Bateman & Fonagy, 1999; Koons et al.,2001; Linehan et al., 1991; 1999).1

Linehan and colleagues (Linehan et al., 1991) compared DBT (Linehan,1993) with community treatment as usual (TAU). They showed that DBTwas generally effective. Compared with TAU, DBT led to a reduction in thenumber and severity of suicide attempts and a decrease in the length of in-patient admissions. In a more recent study (Linehan et al., 1999), DBT wasused to treat drug-dependent women who also have BPD as compared withTAU. DBT patients had more treatment than the TAU patients and they hadsignificantly greater reductions in drug abuse and gains in social adjust-ment. However, in her initial study, there were no between-group differ-ences in the number of hospitalizations or in terms of depression,hopelessness, or reasons for living. Additionally, there were no differencesbetween groups in the number of days hospitalized at 6-month follow up orin self-destructive acts at the end of a 1-year follow up, despite the fact thatthe patients in the DBT group were still receiving DBT therapy, whereas ap-proximately one-half of the TAU group was not in any therapy (Linehan,Heard, & Armstrong, 1993). Whereas the overall results of Linehan’s studyare suggestive of the value of DBT, results from her naturalistic follow up ofpatients in DBT showed variable maintenance of treatment effects, and on-going impairment in functioning in patients who initially experienced symp-tom relief. Although there is understandable enthusiasm for the clinicalusefulness of DBT, more information is needed on the mechanisms anddurability of change effected by DBT (Scheel, 2000; Westen, 2000) and othertreatments for BPD.

In a controlled randomized trial, Bateman and Fonagy (1999) comparedthe effectiveness of 18 months of a psychoanalytically-oriented day hospi-talization program with routine general psychiatric care for patients withBPD. Patients randomly assigned to the day hospital program showed a sta-tistically significant improvement in depressive symptoms and better socialand interpersonal functioning, in addition to a significant decrease in sui-cidal and self-mutilatory acts and the number of inpatient days. AlthoughBateman & Fonagy (2000) showed impressive maintenance of treatment ef-fects in an 18-month follow up, this study lacked a treatment manual andtherapists’ adherence ratings.

LIMITATIONS OF TREATMENT STUDIES

Treatment studies of BPD are few in number, the total number of patientsinvestigated is small, and power is low in each of these studies. Therefore,any generalizations from these studies must be quite tentative in terms ofrelative efficacy of different treatments in relationship to a few domains ofoutcome that have been measured. The outcome domains have been limited

54 CLARKIN ET AL.

1. Other controlled studies reported in the literature are difficult to interpret because of smallpatient group sizes or because the studies focused on either suidical behavior or mixed types ofpersonality disorders without specifying borderline cohorts (Evans et al., 1999; Guthrie et al.,2001; Liberman & Eckman, 1981; Piper, Joyce, McCallum, & Azim, 1998; Salkovskis, Atha, &Storer, 1990).

Page 4: CLARKIN ET AL.BORDERLINE PERSONALITY DISORDER · the personality disorders institute/borderline personality disorder research foundation randomized control trial for borderline personality

and focused mainly on symptoms. The mechanisms of change (mediators ofchange) have rarely been tested, so the evidence for the specific factors in thetreatments that have been investigated is lacking. In addition, both theLinehan et al. (1991) and Bateman and Fonagy (1999) studies did not accessadherence and competence, nor did these studies compare their treatmentagainst another active treatment.

Borderline patients are being identified for treatment studies by the mix-ture of symptoms, attitudes, and behaviors listed as criteria for the disorderin DSM-III and its successors. There are at least two major problems withthe current practice of using DSM-IV diagnoses to select presumed homoge-neous patient groups for treatment intervention: (a) the polythetic diagnos-tic system allows for extensive heterogeneity at the symptom levelrepresented in the diagnostic criteria; and (b) the diagnostic criteria are notstable across time. These surface criteria are variable over time, as some in-dividuals at this level of analysis have the diagnosis at one point in time andnot at another (Zanarini et al., 2003). What remains stable is their relativerank among the group, and their work and social functioning. Further, thesymptom criteria in DSM-IV have an unknown relationship to cognitive,neurocognitive, and affiliative functions of these patients that is likely toguide treatment planning in the future. At the latter, more basic level of de-scription and understanding, borderline patients are characterized by pre-ponderance of negative affect, defective control of affect expression, andconfused and conflicted representation of self and others. It is quite possiblethat the brain functions, neurochemistry, and neurocognitive functioning ofthese patients will provide more important ways to classify these patients,and to identify both targets for treatment and subgroups of patients for morespecific treatments.

As long as the DSM-IV criteria are used to select patient groups for empiri-cal treatment research without supplemental descriptions of the patients,the efforts to find a clear relationship between a defined intervention and itseffects on a homogeneous group of patients with clear goals for treatmentwill be compromised, if not totally obscured. A central question is which ofthe constructs that are heterogeneous among these patients is crucial totreatment elements and long-term effects.

In summary, psychotherapy research focused entirely on group meanscores before and after a specified treatment yield little information. Psycho-therapy research can be a method of empirically examining and teasing outthe various elements and aspects of a particular disorder. In turn, by sepa-rating out the elements of a disorder, treatment can be developed to targetthe various domains of the disorder that may respond differentially to differ-ent treatments. In reviewing existing treatment research on personality dis-orders, Crits-Christoph, Cooper, and Luborsky (1998) stated that futureresearch must match the focus of the treatment to the nature of the disor-der, and provide measures of change across time. The form and content ofpsychotherapy for individuals with personality disorders in general andwith BPD more specifically is in a state of treatment development ratherthan consolidation.

BORDERLINE PERSONALITY DISORDER 55

Page 5: CLARKIN ET AL.BORDERLINE PERSONALITY DISORDER · the personality disorders institute/borderline personality disorder research foundation randomized control trial for borderline personality

WORKING MODEL OF BPDCentral to Kernberg’s conceptualization of borderline personality organiza-tion (TFP model of treatment) are mental representations that are derivedthrough the internalization of attachment relationships with caregivers. Thedegree of differentiation and integration of these representations of self andother, along with their affective valence constitutes personality organization(Kernberg, 1984). Borderline personality can be thought of as a severely dis-turbed level of personality organization, characterized by the use of primi-tive defenses (e.g., splitting, projective identification, dissociation), identitydiffusion (e.g., inconsistent view of self and others), and deficits in realitytesting (e.g., poor conception of one’s own social stimulus value).

Our current working model of BPD is an incorporation of thepsychodynamic conceptualization of Kernberg concerning identity diffu-sion, and the more empirically grounded conceptions of negative affect,self-regulation (constraint and effortful control), and affiliative bonding(Clarkin & Posner, in press; Posner et al, 2003; Depue & Lenzenweger,2001). This conceptualization of the disorder and its major componentsinforms baseline data collection and areas of assessed change in treatment.

We do not assume that a temperamental disposition of negative affect andpoor effortful control will result in BPD. Rather, it is assumed that thesetemperamental dispositions in the context of an environment involving earlyseparations, physical or sexual abuse, and parental neglect can lead toidentity diffusion and impulsive, self-destructive behavior. Otherneurobehavioral systems could also interact with the basic high negative af-fect/low control (constraint) to potentiate the expression of a BPD-pronetemperament (Depue & Lenzenweger, 2001).

AFFECT

Negative affect, especially hostility and aggression, with relatively minimalpositive affect is an essential aspect in understanding the individual withBPD (Kernberg, 1984; Depue & Lenzenweger, 2001). Negative affect invadesthe information processing of the individual (Silbersweig et al., 2001) andthe organization of the individual’s interpersonal and personal experience.

SELF-REGULATION

A second central feature of borderline pathology is poor self-regulation. Thisrelative inability to self-regulate is manifested in impulsive behaviors, in-cluding impulsive self-destructive behaviors, and difficulties in regulatingaffect. The construct of impulsivity has been defined differently in a varietyof studies, involving the following elements: (a) rapid, unplanned reactionsto stimuli; (b) decreased sensitivity to negative consequences of behavior;and (c) lack of regard for long-term consequences of behavior (Moeller,Barratt, Dougherty, Schmitz, & Swann, 2001).

Impulsivity or impulsive aggression are considered to be underlying di-mensions in BPD (Siever & Davis, 1991; Zanarini, 1993; Links, Heslegrave,& van Reekum, 1999). In a stepwise, multiple-regression model, the im-pulse action score from the Diagnostic Interview for Borderline Personality

56 CLARKIN ET AL.

Page 6: CLARKIN ET AL.BORDERLINE PERSONALITY DISORDER · the personality disorders institute/borderline personality disorder research foundation randomized control trial for borderline personality

Disorder (DIB) best predicted borderline psychopathology at follow up(Links et al., 1999). Impulsivity combined with other factors has been re-lated to suicidal behavior in BPD patients. For example, Soloff, Lynch, andKelly (2002) found impulsive actions, comorbid antisocial personality disor-der, and depression related to a history of suicidal behavior in BPD patients.In a diagnostically mixed group of patients, aggression and impulsivity werehigher in suicide attempters, compared with those without suicide attempts(Mann, Waternaux, Haas, & Malone,1999).

There is evidence of the link between impulsivity and underlying biologicalsystems (Depue & Lenzenweger, 2001). Both impulsive aggression and af-fective instability show a stronger familial relationship than the diagnosis ofBPD itself (Silverman et al., 1991). In twins, impulsivity and affective insta-bility are heritable (Torgersen, 1984; Torgersen et al., 2000). Biological,neuroendocrine, and imaging studies provide evidence for the involvementof serotonergic activity in impulsive aggression (Coccaro et al., 1989; Siever& Trestman, 1993; Gurvits et al., 2000).

Affect dysregulation or emotional instability has been described as involv-ing unpredictability of responses to stimuli, increased lability of baseline,unusual intensity of responses, and unusual responses (Spoont, 1996), allcharacteristics of a poorly constrained biobehavioral regulatory system(Mandell, Knapp, Ehlers, & Russo, 1984; Spoont, 1992). Patients with affec-tive disorders display dysregulation of positive affectivity (Depue & Spoont,1986; Spoont, 1992), whereas BPD patients have explicit dysregulation ofnegative affect (Spoont, 1996). It is hypothesized that BPD patients have ele-vated levels of negative affect in conjunction with a nonaffective constraintsystem that is less effective (Depue & Lenzenweger, 2001; Spoont, 1996).

The evolution of self-regulation in the developing child is central to under-standing both the development of normal personality and its organizationand personality pathology (Posner & Rothbart, 2000). Studies suggest thateffortful control has a developmental course in which some children by age 3years are capable of using executive control systems to efficiently makechoices in conflict situations, especially those involving the suppression ofdominant response modes. Effortful control is related to approach of situa-tions that involve aversive stimuli and avoidance of situations that may giveimmediate reward. This capacity of inhibiting a predominant response in fa-vor of a subdominant one is considered a form of behavioral self-control,and, therefore, a mechanism of self-regulation, in that the individual has theability to control arousal and response.

The constructs of arousal, affect, and self-regulation are central to a num-ber of conceptualizations of human psychological development (Derryberry& Rothbart, 1988). Both motivational systems (including appetitive and ap-proach behavior, fearful behavior, frustrative and aggressive behavior, andaffiliative and nurturing behavior) and attentional systems are seen as con-tributors to the psychological development of the individual. For example,the shifting of attention from a negative to a positive stimulus can helpsoothe a distressed child (Derryberry & Rothbart, 1988). The critical role ofeffortful control in socialization is reflected in research showing thateffortful control is positively related to conscience development (Kochanska,

BORDERLINE PERSONALITY DISORDER 57

Page 7: CLARKIN ET AL.BORDERLINE PERSONALITY DISORDER · the personality disorders institute/borderline personality disorder research foundation randomized control trial for borderline personality

1995; Kochanska, Murray & Coy, 1997) and negatively related to theexpression of aggression (Rothbart, Ahadi, & Hershey, 1994).

EVOLVING CONCEPTUALIZATION OF SELF AND OTHERS

Many (Bowlby, 1969, 1973,1980; Kernberg, 1996) have postulated that thedeveloping child evolves a working conceptualization of self and others, es-pecially under the influence of affectively charged interactions with othersthat are comforting and pleasurable or aversive and dangerous. It is fromthese early interactions that the developing individual builds an internalmodel of self and others that subsequently provides expectations in laterinteractions with others.

Influenced by temperamental disposition, environmental (traumatic)events or a combination of both, a secondary level of intrapsychic organiza-tion takes place that determines the clinical syndrome of identity diffusion(Kernberg, 1996) that is reflected in the DSM-IV diagnostic criteria for BPD.Identity diffusion is characterized by a lack of integration of the concept ofself and the related concept of significant others. These poorly integratedconceptions of self and others are derived from an excessive splitting, oftenreferred to as dichotomous thinking, or primitive dissociation between posi-tive and negative affective investment of self and other representations,leading to the chronic deficiency in the assessment of self and self-motiva-tions. The clinical characteristics of identity diffusion are chronic immatu-rity in judgments of emotional relationships, difficulties in the commitmentto intimate relations and disturbances in sexual and love life, and problemswith commitment to work or to a profession.

Recently, clinical researchers and theorists have understood fundamen-tal aspects of BPD such as unstable, intense interpersonal relationships,feelings of emptiness, bursts of rage, chronic fears of abandonment and in-tolerance for aloneness, as stemming from impairments in the underlyingattachment organization (Blatt, 1995; Fonagy et al., 1996; Gunderson,1996).

TRANSFERENCE-FOCUSED PSYCHOTHERAPYAmong several other promising treatment approaches to BPD is the objectrelations approach based on Kernberg’s clinical theorizing (Kernberg, 1984;1996). Kernberg and colleagues call this treatment transference-focusedpsychotherapy (TFP; Clarkin et al., 1999) because it relies principally on thetechniques of clarification, confrontation, and interpretation within theevolving transference relationship between the patient and the therapist.With the assistance of an NIMH treatment development grant (awarded toJohn Clarkin), we have provided evidence that TFP is effective using patientsas their own controls (Clarkin et al., 2001), and in comparison to a TAU BPDgroup (Levy, Clarkin, Foelsch, & Kernberg, 2003).

The major goals of TFP are better behavioral control, increased affect regu-lation, more intimate and gratifying relationships and the ability to purse lifegoals. This is believed to be accomplished through the development of inte-grated representations of self and others, the modification of primitive de-

58 CLARKIN ET AL.

Page 8: CLARKIN ET AL.BORDERLINE PERSONALITY DISORDER · the personality disorders institute/borderline personality disorder research foundation randomized control trial for borderline personality

fensive operations and the resolution of identity diffusion that perpetuatethe fragmentation of the patient’s internal representational world. In thistreatment, the analysis of the transference is the primary vehicle for thetransformation of primitive (e.g., split, polarized) to advanced (e.g., complex,differentiated, and integrated) object relations.

TFP begins with explicit contract setting that clarifies the conditions oftherapy, the method of treatment, and the respective roles of patient andtherapist. The primary focus of TFP is on the dominant affect-laden themesthat emerge in the relationship between borderline patients and their thera-pists in the here and now of the transference. During the first year of treat-ment, TFP focuses on a hierarchy of issues: (a) the containment of suicidaland self-destructive behaviors; (b) the various ways of destroying the treat-ment; and (c) the identification and recapitulation of dominant object rela-tional patterns, as they are experienced and expressed in the here and nowof the transference relationship.

Within psychoanalysis, the TFP approach is closest to the Kleinian school(Steiner, 1993), which also emphasizes a focus on the analysis of the trans-ference. However, TFP can be distinguished from Kleinian psychoanalysis inthat in that TFP is practiced twice per week and includes a more highlystructured treatment frame by emphasizing the treatment contract and anestablished set of priorities on which to focus (e.g., suicidality, treatment in-terfering behaviors, etc.). The role of the treatment contract and the treat-ment priorities both go beyond that found in more typical psychoanalyticpsychotherapy or psychoanalysis, including Kleinian psychoanalysis. Inaddition, transference interpretations are consistently linked with bothextratransference material and, importantly, long-term treatment goals(e.g., better behavioral control). Although TFP adheres more strictly to tech-nical neutrality than many psychodynamic treatments (Buie & Adler, 1982;Waldinger & Gunderson, 1989), in contrast to Kleinian approaches, the TFPapproach is a highly engaged, more talkative, and interactive. Additionally,technical neutrality is de-emphasized to the extent required to maintainstructure. TFP also differs from other expressive psychodynamic ap-proaches with a persistent focus on the here and now, a focus on the imme-diate interpretation of the negative transference, and an emphasis oninterpretation of the defensive function of idealization, as well as a focus onthe patients’ aggression and hostility. In contrast, many otherpsychodynamic approaches view the central task, particularly of the earlyphase, as primarily supportive and relationship building (Buie & Adler,1982; Chessick, 1979; Masterson, 1981) and foster idealizing aspects of thetransference (Buie & Adler, 1982; Chessick, 1979). Some recent articula-tions of psychodynamic approaches de-emphasize working in the symbolicrealm, noting that the BPD patient’s ability to understand and use interpre-tations varies widely, and instead stress supportive, behavioral, orpsycho-educational techniques (Rockland, 1992; Gunderson, Berkowitz,Ruiz-Sancho, 1997; Gunderson & Wheelis, 1999).

In relation to DBT, some of the most salient differences between the twotreatments center on the frame. The TFP therapist is considered unavailablebetween sessions, whereas in DBT the patient is encouraged to phone theindividual therapist between sessions. Another difference is the TFP empha-

BORDERLINE PERSONALITY DISORDER 59

Page 9: CLARKIN ET AL.BORDERLINE PERSONALITY DISORDER · the personality disorders institute/borderline personality disorder research foundation randomized control trial for borderline personality

sis on technical neutrality versus the DBT validation, coaching, and cheer-leading strategies. Despite these differences, both TFP and DBT have incommon a firm, explicit contract, a focus on a hierarchy of acting out behav-iors, a highly engaged therapeutic relationship, and a structured disciplinedapproach.

In terms of patient selection, Kernberg (1984) suggests that borderline pa-tients with narcissistic, paranoid, and antisocial personality disorder,termed malignant narcissism, would be more difficult to engage in treatmentand have a poorer prognosis. Others have suggested that TFP is designedonly for those patients with sufficient motivation, intelligence, and psycho-logical mindedness (Swenson, 1989). Further research is needed in under-standing patient prognostic factors for TFP. In the current study patientswere not assessed for inclusion based on any of these variables or any othersigns of good prognosis, nor were patients excluded because of indicators ofpoor prognosis.

How does TFP relate to the model of BPD with its emphasis on the con-structs of negative affect, effortful control and constraint, and conception ofself and others? TFP focuses its therapeutic efforts on the relationship be-tween therapist and patient in which the patient demonstrates by words,feelings, and actions his or her conceptualization of the other (therapist) inrelationship to conceptualization of self with associated affects. Through theclarification and understanding of the expectations and distortions that thepatient brings to that relationship, it is thought that the patient slowlyevolves a more integrated sense of self and realization that expectations ofothers based on prior experiences are often inaccurate. It is hypothesizedthat through the more coherent conceptualization of self and others, the pa-tient achieves control over affects, especially as they are related tointerpersonal interactions.

ORGANIZATION OF OUR RESEARCH EFFORTOur psychotherapy study is used as a platform on which we recruit, assess,and provide treatment for patients with BPD. On this basic structure, wehave built a procedure that provides data on pathology and neurocognitivefunctioning. The primary purpose of the study is to examine the efficacy ofthree standard treatments for BPD of 1-year duration. First, a cognitive-be-havioral treatment called DBT (Linehan, 1993), was compared with apsychodynamic treatment called TFP (Clarkin et al., 1999) and with a sup-portive treatment (Rockland, 1992), used to contrast with these two activetreatments as a control for attention and support. Secondary and tertiarypurposes were: (a) to create a database of well-characterized patients in or-der to examine patient personality and neurocognitive variables that predicttreatment response as well as the subsequent naturalistic course of the dis-order; and (b) to examine the processes and mechanisms of change duringthe course of psychotherapy.

This treatment study of BPD patients is unique and goes beyond existingtreatment studies in a number of ways: (a) this is the first BPD treatmentstudy to include males; (b) this study includes not only borderlines with sui-cidal behavior, but all participants who meet the diagnosis; (c) this is the

60 CLARKIN ET AL.

Page 10: CLARKIN ET AL.BORDERLINE PERSONALITY DISORDER · the personality disorders institute/borderline personality disorder research foundation randomized control trial for borderline personality

first study to compare two forms of active treatment to a supportive treat-ment; (d) therapists are not located at a university clinic or hospital but intheir private offices in the community; (e) medication is carefully delivered,when needed, by an algorithm; patients with and without medication pro-vide a contrast in the data analysis; and (f) outcome measures involve notonly symptom change, but also changes in organization of the personality atthe psychological and neurocognitive levels.

PROCEDURE

Patients were assessed with a number of semistructured interviews andself-report instruments to establish the diagnosis. Diagnostic instrumentsincluded the Structured Clinical Interview for DSM-IV Axis I Disorders(SCID-I; First, Gibbon, Spitzer, & Williams, 1996) and the International Per-sonality Disorder Examination (IPDE; Loranger, 1999). Assessment instru-ments were chosen to reflect important domains that might show change intreatment such as symptoms, behaviors, attention, positive and negative af-fect, affect regulation, work and social functioning, identity, and identitydiffusion.

Patients were also assessed using neurocognitive tasks known to tap ex-ecutive functioning and attention (e.g., the Attention Network Task, the Wis-consin Card Sort Test, the Continuous Performance Test). A subgroup ofpatients underwent functional imaging (fMRI) before and after 1 year oftreatment. On completing the assessment patients were randomized to oneof the three treatment conditions for 1-year outpatient treatment.

Patients. The BPD patients were recruited from New York City and adja-cent Westchester County, referred by private practitioners, clinics, familymembers and self-referred. To foster subject retention over the duration ofthe study, we recruited patients who lived within a 50-mile radius of thestudy site. Participants were males and females between the ages of 18 and50 years. Patients with comorbid schizophrenia, schizoaffective disorder,bipolar disorder, delusional disorder, delirium, dementia, amnestic andother cognitive disorders were excluded because of the influence of brain pa-thology and thought disorder on the ability to provide meaningful self-reportdata and complicated response to treatment. The exclusion of psychotic pa-tients from a study of personality disorder is a common research standard atmany centers and allows for a “cleaner window” on personality pathology.We include patients with other comorbid Axis I disorders, as issues ofcomorbidity across time with these Axis I disorders is a focus ofinvestigation.

We were clinically referred and interviewed 207 individuals for at leastone evaluation session. Of these 207 participants, 109 were eligible forrandomization. Most exclusions were due to the absence of five criteria forBPD (N = 34). The second most common reason was age (N = 30), followedby nine patients who met criteria for current substance dependence, eightpatients who met criteria for schizophrenia or a schizophrenic disorder,eight patients who dropped out of the evaluation process, six patients whomet criteria for bipolar I disorder, two patients who had IQs lower than 80,and one patient who had a scheduling conflict. Of the 109 patients eligible

BORDERLINE PERSONALITY DISORDER 61

Page 11: CLARKIN ET AL.BORDERLINE PERSONALITY DISORDER · the personality disorders institute/borderline personality disorder research foundation randomized control trial for borderline personality

for randomization, 90 were randomized to treatment. There were no differ-ences in terms of demographics, diagnostic data, and severity ofpsychopathology between those patients randomized to treatment andthose that were not.

TREATMENT AND THERAPISTS

Patients were randomized to one of the three treatment conditions for 1-yearoutpatient treatment. Two treatments, a cognitive-behavioral treatmentcalled DBT (Linehan, 1993), and a psychodynamic treatment called TFP(Clarkin et al., 1999), have received preliminary empirical support for theireffectiveness. The mechanisms of change in these two treatments are con-ceived in very different ways. DBT is hypothesized to operate through thelearning of emotion regulation skills in the validating environment of thetreatment (Linehan, 1993). TFP is hypothesized to operate through the inte-gration of conflicted, affect-laden conceptions of self and others via the un-derstanding of these working models as they are actualized in thehere-and-now relationship with the therapist. A third treatment, called sup-portive treatment (Rockland, 1992), is used in contrast to these two activetreatments as a control for attention and support.

Therapists in each of the three treatment conditions were selected basedon prior demonstration of competence in the treatment. In order to ensureongoing therapist adherence and competence, all treatments were super-vised on a weekly basis by experts in each treatment. Barbara Stanley, PhD,an acknowledged expert in DBT and NIMH funded researcher in this area, isthe supervisor of DBT. Otto Kernberg, a psychoanalyst of international stat-ure, is the supervisor of TFP. Ann Appelbaum, expert therapist, is thesupervisor of the supportive treatment.

MONITORING OF TFP, DBT, AND SPT

Treatment integrity was monitored in a number of ways. First, each treat-ment cell leader was responsible for recruiting therapists for their treatmentcell. Therapists were known to the treatment cell leaders. Second, therapistsin each treatment cell attended weekly group supervisions where treatmentcell leaders were able to observe videotaped sessions. Feedback to thera-pists was provided by treatment cell leaders whenever a therapist fell belowan acceptable level of either adherence to the manual or competence. Whena therapist’s ratings were consistently low for adherence, then ratings weremade more frequently (approximately every 4 sessions) for the succeeding3-month interval, and supervision focused on the difficulties identified byraters. Additional supervision was provided when either adherence or com-petence fell below acceptable levels. When a therapist fell below acceptablelevels, no new cases were assigned to them. Third, we asked treatment cellleaders to rate and rank the therapist on each case. Finally, both expert rat-ers and independent naive raters evaluated videotaped TFP, DBT, and SPTsessions for adherence and competence at regular intervals. Raters as-sessed therapists for adherence and competence on every 10 sessions, over

62 CLARKIN ET AL.

Page 12: CLARKIN ET AL.BORDERLINE PERSONALITY DISORDER · the personality disorders institute/borderline personality disorder research foundation randomized control trial for borderline personality

two consecutive sessions, beginning with session number 10. Adherenceand competence rating were averaged over these sessions.

DOMAINS OF OUTCOME

The domains of outcome in a psychotherapy study are determined by thegoal of the treatment (e.g., what patient changes does the therapy intend)and the hypothesized mechanisms of change (e.g., predictors, mediators,and moderators). Thus, in our ongoing treatment study we assess the influ-ence of treatment in reference to the central temperamental features of neg-ative affect (i.e., lowered negative affect) and effortful control (i.e., increasedeffortful control/constraint), in addition to the changes in the BPD Axis IIcriteria themselves. The advantage of assessing change in these two keytemperamental dimensions is their close relationship to underlyingneurobehavioral systems of the organism on the one hand, and their obvi-ous impact on everyday functioning on the other. We postulate that de-crease in negative affect (or change in the balance of positive and negativeaffect) and increase in effortful control would be features of any successfultreatment of BPD patients. Focus on these variables provides a context inwhich we can judge the relative success of different types of psychosocialtreatment. It also provides us with a unique opportunity to determine ifthere are specific gains that maintain or accrue with each of the three treat-ments. For example, there may be notable gains associated with one of thethree treatments in one area of psychosocial functioning, which are them-selves seen in correlation with changes in negative affect or effortful control.We also assess the third crucial variable of conceptualization of self andothers, as it is through the conceptualization of the interpersonal world thatthe individual controls and modulates affect.

PLANNED STATISTICAL ANALYSES: PRECISION AND SENSITIVITYFOR DETECTING CHANGEMost psychotherapy studies are underpowered (especially those that com-pare one or more active treatments), do not have focused predictions regard-ing outcome, and use techniques (repeated-measures multivariate analysisof variance [MANOVA]) that are inappropriate for detecting change inmultiwave data. Given the limitations of the amount of funding provided bythe BPDRF for this initial study of treatments in BPD, we developed a studyprotocol that would contrast three treatments in 90 patients across time (12months). Our design, therefore, included the use of multiple data collectionpoints and data-analytic procedures that would maximize the power to de-tect change in our patients, in relation to the treatment modalities tested.

The variables of primary interest in this psychotherapy outcome study arecontinuous in nature and each of these variables was assessed at four timepoints, namely at baseline, 4 months, 8 months, and 12 months (termina-tion of treatment). Thus, each study subject will have been measured on thesame variables at roughly the same intervals at four points in time. The datafor this study will be analyzed principally from two different perspectives.The first perspective, which represents a more traditional (pre- vs.

BORDERLINE PERSONALITY DISORDER 63

Page 13: CLARKIN ET AL.BORDERLINE PERSONALITY DISORDER · the personality disorders institute/borderline personality disorder research foundation randomized control trial for borderline personality

posttreatment) approach with the added precision of contrast analysis, willinvolve the evaluation of treatment gain scores, which will simply be theamount of improvement observed from baseline to the 12-month assess-ment on a given dimension of interest (e.g., BPD symptoms, negative affect,impulsivity, self-destructive behaviors, etc.). These “gain scores,” which willbe adjusted for initial level on the variable of interest at baseline, will then becompared across the three treatment modalities through a series of focusedcontrast analyses within an analysis of variance (ANOVA) framework(Rosenthal, Rosnow, & Rubin, 2000; Rosenthal & Rosnow, 1991). Our dataconfiguration is ideally suited for the a priori specification of theory-guidedcontrast analyses. For these one-way ANOVA analyses (with focused con-trasts), the patients will be subdivided across a between-subjects (B-S) fac-tor with 3 levels (3 types of treatment). Thus, the one-way ANOVA will turninto what is essentially a single degree of freedom test (in the numerator),tested with the t-statistic, and yielding an “effect-size r.” It is noted thatrather than using a traditional “unfocused” approach to ANOVA, the fo-cused contrast analysis approach reduces the degrees of freedom, whichhas the net effect of both a more powerful and more precise analysis(Rosenthal & Rosnow, 1991; Rosenthal et al., 2000). Power analyses suggestthat, within a contrast analysis framework, the present study should haveadequate power (80% or more) to detect modest effects, even with someattrition in the sample.

The second approach to the analysis of change in the dimensions of inter-est for the patients in this study will involve the application ofstate-of-the-art individual growth curve analysis. Analysis of individualgrowth curves will be done via a multilevel modeling approach (Goldstein,1995) (also known by some as hierarchical linear modeling [Raudenbush &Bryk, 2001]). In this powerful statistical framework, which is ideally suitedto multiwave data, we will investigate our continuous dependent variablesin a model that casts them as Level I variables organized by time (i.e., nestedwithin persons) and Level II variables (between persons) which include thenesting variables, such as treatment group and sex. The individual growthanalyses for the data from the patients will proceed in a sequential fashion:(a) estimation of unconditional growth models with intervals between as-sessments, wherein time is defined from study entry and intervals (months)will be centered on the individuals’ means; (b) estimation of a Level II modelthat includes age at entry to study in the prediction of initial status (inter-cept, “elevation”) and change (slope, “change”) values retained from Level I;and (c) estimation of a Level II model that includes age at entry, sex, andtreatment group. Additional Level II models will also be estimated after thesebasic models and they shall typically include baseline measures (e.g., tem-perament constructs) as time-invariant predictors at Level II. The four as-sessment points that we will have for each of the dependent variables ofinterest will allow us to fit not only linear growth models, but also quadraticmodels. The multilevel modeling approach is statistically superior to a re-peated-measures MANOVA (and ANOVA) approach that is now regarded asoutmoded by most longitudinal methodologists (Singer & Willett, 2003). Themultilevel modeling approach can handle missing data very effectively anddoes not require that all subjects have complete data at all assessment

64 CLARKIN ET AL.

Page 14: CLARKIN ET AL.BORDERLINE PERSONALITY DISORDER · the personality disorders institute/borderline personality disorder research foundation randomized control trial for borderline personality

waves (unlike MANOVA). The individual growth curve analytic approachhas recently been applied to longitudinal data on personality disorders andthis analysis yielded insights into the stability and change of PD featuresover time in a manner that could not have been achieved with repeatedmeasures MANOVA (Lenzenweger, Johnson, & Willett, submitted).

PRELIMINARY DATA ON PATIENT CHARACTERISTICS

We review the preliminary data on patient demographics, the crucial areasof negative affect, affect modulation, conception of self and others, and envi-ronmental variables. These theoretical cruicial patient variables will be im-portant correlates of patient behavior in the randomized treatment study.

The patients were predominantly female (92%), with a mean age of 31years. They were ethnically diverse and patients were 62% Caucasian, 10%African American, 9% Hispanic, 5% Asian, and 8% other. They had first con-tact with psychiatric treatment at a mean age of 17 years. The mean GlobalAssessment of Functioning (GAF) score at the time of admission into thestudy was 50, indicating a substantial degree of symptoms and disruptedfunctioning. Whereas all patients met criteria for BPD, they were heteroge-neous in terms of co-existing personality disorders and Axis I conditions. Interms of suicidal behavior, 57% manifested prior suicidal behavior, 64%manifested prior parasuicidal behavior, and 17% (N = 15) had a history ofneither.

CHALLENGES ENCOUNTERED IN THE DESIGN AND EXECUTION OFA THERAPY STUDY OF BPDIn any empirical study of psychotherapy there are compromises that divertfrom an ideal design. We think it is helpful to review some of the design is-sues that we faced, as others could confront them and may profit from ourexperience.

First, the three treatments were delivered with attention to preserving theintegrity of each treatment under investigation (Elkin, Pilkonis, Docherty, &Sotsky, 1988). Thus, the frequency of contact between patient and thera-pist, the nature of the treatment, and the manner in which the treatmentwas ended was not strictly equalized across the three treatment conditions,but rather delivered in accordance with the specific definition of each of thethree treatments.

We are aware, therefore, that the number of hours of contact between pa-tients and therapists varies between the three treatment conditions. Sup-portive treatment is one 50-minute contact per week, TFP is two 50-minutecontacts per week, and DBT is 1 hour of individual therapy contact and oneand one-half hours group treatment contact per week. In addition, the pa-tients in DBT are encouraged to telephone the therapist between sessions.We have tracked the contact time in each treatment condition. However, at-tempts to equalize the contact time between the three treatment conditionswould threaten the integrity of the treatments. We delivered the treatmentsas designed, and will do a cost analysis of the treatments to compare the costto the benefit ratio of each.

BORDERLINE PERSONALITY DISORDER 65

Page 15: CLARKIN ET AL.BORDERLINE PERSONALITY DISORDER · the personality disorders institute/borderline personality disorder research foundation randomized control trial for borderline personality

Second, allegiance effects were a design issue. Those researchers with apassionate interest and dedication to a particular approach to therapy arethe likely ones to have the energy and enthusiasm to investigate that treat-ment in some form of randomized clinical trial. This is both humanly neces-sary and a problem in the interpretation of the outcome of such studies(Luborsky et al., 1999).

We attempted to control potential allegiance effects by several ways. First,we placed randomization to treatment in the hands of a researcher in our de-partment who was independent of the study and not informed about ourstudy hypotheses. Secondly, patients were not treated within the physicalenvironment of the medical school setting that was the site of assessmentand randomization. Rather, patients were treated in the private offices oftherapists who were committed to one of the three treatments under study.

Third, the manner in which patients are recruited, and the way they arehandled between recruitment and the initiation of the treatment can havesignificant effects on the results of the study, including dropout rate and thegeneralizability of the results.

Fourth, it is common in randomized clinical trials of the treatment ofsymptom disorders to have as the primary outcome measure symptoms thatare the target of change. Often, the outcome domains are not extensive andare mainly related to symptoms. Even in the treatment of personality disor-ders, such as BPD, the main domains of measurement involve symptomssuch as suicidal behavior and depression (Linehan et al., 1991). It is cus-tomary for statistical reasons to state a few primary areas of change in thedesign. Whereas this design issue is understandable, we think that the do-mains of change must be expanded to understand the nature of the pathol-ogy in interaction with the nature of the treatment, and to understand longterm benefits of the treatments. As described earlier in this article, ourmodel of borderline personality pathology suggests that negative affect,effortful control, and conceptualization of self and others are crucial ele-ments in the pathology, and, therefore, foci for change in treatment. Thisconceptualization would indicate that these constructs be measured preand post treatment.

Because different treatments have diverse treatment foci, while at thesame time attempting to change the same symptom complex, we thought itwould be important to measure constructs seen as mechanisms of changeor mediating variables in the treatments, to be measured along with thesymptom outcomes. In the context of validation, DBT posits the education ofthe patient in the use of skills that should lead to the reduction of suicidalbehavior, control of affect, and more prosocial behavior. TFP attempts tomodify suicidal and self-destructive behavior through the modification ofinternalized conceptions of self and others that are hypothesized to be polar-ized and dysfunctional in the borderline patient. In short, we anticipate thatdifferent treatments will affect different domains of functioning in these BPDpatients; we do not expect all three treatment modalities to affect all aspectsof BPD symptomatology, personality functioning, and behavior outcomescomparably.

Fifth, it is often assumed that the type of treatment provides the specific ef-fect, and that therapists delivering a specific type of treatment are adherent

66 CLARKIN ET AL.

Page 16: CLARKIN ET AL.BORDERLINE PERSONALITY DISORDER · the personality disorders institute/borderline personality disorder research foundation randomized control trial for borderline personality

and competent, and thus do not create a variable influencing results. Thisassumption of therapist uniformity has been questioned and most recently,Wampold (2001) has made extensive use of meta-analysis to demonstratethat therapist effects are more potent than brand of therapy effects.

Our approach was threefold: (a) to control for therapist adherence andcompetence by training before the initiation of the study; (b) ongoing super-vision of all therapists during the study to ensure timely adherence andcompetence; and (c) sufficient numbers of patients per therapist to examinethe data for therapist effects.

Sixth, we discussed in some detail the exact quantification of “one year” oftreatment. Because there can be legitimate reasons to “suspend” treatment(e.g., patient or therapist medical illness) during the year following admis-sion to the study, it was decided to define the treatment as 50 weeks of treat-ment exposure that could take place over a time period of up to 13.5 months.At the end of 1 year of treatment, the treatment (i.e., medication treatmentand either supportive, TFP, or DBT) provided free of charge to the patient willend. The controlled part of the study will end at that time, and at that pointthe naturalistic follow up period will begins.

Seventh, borderline patients are extremely sensitive to issues of attach-ment and feared abandonment, and this applies to their treatment provid-ers and to significant others in the environment. In order to control for thepatients’ mindset regarding the availability of their study therapist, patientsin all three treatment conditions were informed of the 1-year duration oftreatment financed by the grant and free of charge to the patient. In this way,patients in all treatment conditions were faced with the dilemma of 1 year ofpredicable attachment to a therapist that would be potentially changed at aknown date. In this way, we hope to disentangle the effects of the specifictreatments, from the anticipation of duration of attachment to a knowntherapist.

Because these are very disturbed patients, it is not expected that theirtreatment will be completed at 1 year. However, the reality of the currenthealth care system is that 1 year of treatment is more than what can usuallybe obtained. The rationale for follow up is to assess the maintenance of treat-ment gains. This will be a naturalistic follow up because some patients willcontinue with the study therapists, others will seek treatment elsewhere,and others will receive no further treatment.

Eighth, because all patients were provided with an evaluation by a psychi-atrist and the prescription of medication if deemed appropriate by instru-ments and guidelines, we have a design in which some patients receive acombination of medication and one of three types of psychosocial treatment,and some patients receive no medication because it is deemed unnecessary.This creates a complicated research situation.

We considered several options. First, we could have included only border-line patients in the study that did not meet indications for medication, andcompared three forms of treatment for those selected patients. We thoughtthe price to pay for this design would be lack of external validity. A study ofborderline patients limited only to those not in need of medication would notbe representative of the majority of borderline patients. Another design al-ternative would have been to select a subset of borderline patients who

BORDERLINE PERSONALITY DISORDER 67

Page 17: CLARKIN ET AL.BORDERLINE PERSONALITY DISORDER · the personality disorders institute/borderline personality disorder research foundation randomized control trial for borderline personality

could be adequately treated with one class of medication (e.g., those whocould be treated with a particular SSRI). This option, too, would havereduced generalizability of the study results.

THE NEAR FUTURE: TREATMENT DEVELOPMENTAs the phenomenology and mechanisms of borderline personality pathologybecome clearer, it provides an opportunity to either modify existing ap-proaches to the specifics of the pathology and/or important subgroups, or todevelop new treatments that address the issues. In our on-going random-ized clinical trial, we will utilize contrast analyses to explore specific hypoth-eses concerning the relative efficacy of the three treatments in question,given our current understanding of the outstanding features of the patients’pathology. These findings may provide leads as to which treatments havethe most impact on patient subgroups in terms of specific domains ofchange. All changes that appear after one year of treatment must beassessed for their durability over time in a longer-term follow up.

68 CLARKIN ET AL.

REFERENCES

Bateman, A., & Fonagy, P. (1999). Effective-ness of partial hospitalization in thetreatment of borderline personality dis-order: A randomized controlled trial.American Journal of Psychiatry,156,1563-1569.

Bateman A.W., & Fonagy, P. (2000). Effective-ness of psychotherapeutic treatment ofpersonality disorder. British Journal ofPsychiatry, 177, 138-143.

Blatt, S. J. (1995). Representational struc-tures in psychopathology. In D.Cicchetti, & S. Toth (Eds.), RochesterSymposium on DevelopmentalPsychopathology: Vol. 6. Emotion, Cog-nition and Representation (pp. 1-33).Rochester, NY: University of RochesterPress.

Blum, N., Pfohl, B., St. John, D., Monahan,P., & Black, D. W. (2002). STEPPS: Acognitive-behavioral systems-basedgroup treatment for outpatients withborderline personality disorder—a pre-liminary report. Comprehensive Psychi-atry, 43, 301-310.

Bowlby, J. (1969/1982). Attachment and loss:Vol. 1. Attachment. New York: BasicBooks.

Bowlby, J. (1973). Attachment and loss: Vol. 2.Separation. New York: Basic Books.

Bowlby, J. (1980). Attachment and loss: Vol. 3.Loss. New York: Basic Books.

Buie, D.H., & Adler, G. (1982). Definitivetreatment of the borderline personality.International Journal of Psychotherapy,9, 51-87.

Chessick, R.D. (1979). A practical approachto the psychotherapy of the borderlinepatient. American Journal of Psychia-try, 33, 531-546.

Clarkin, J. F. (1996). Treatment of personalitydisorders. British Journal of ClinicalPsychology, 35(4), 641-642.

Clarkin, J. F., Foelsch, P. A ., Levy, K. N., Hull,J. W., Delaney, J. C., & Kernberg, O. F.(2001). The development of apsychodynamic treatment for patientswith borderline personality disorders:A preliminary study of behavioralchange. Journal of Personality Disor-ders, 15(6), 487-495.

Clarkin, J. F., & Posner, M. (in press). Defin-ing the mechanisms of borderline per-sonality disorder. Psychopathology.

Clarkin, J.F., Widiger, T., Frances, A., Hurt,S.W., & Gilmore, M. (1983). Prototypictypology and the borderline personalitydisorder. Journal of Abnormal Psychol-ogy, 92, 263-275.

Clarkin, J. F., Yeomans, F., & Kernberg, O. F.(1999). Psychotherapy of borderlinepersonality. New York: Wiley.

Coccaro, E. F., Siever, L. J., Klar, H., Maurer,G., Cochrane, K., Cooper, T. B., Mohs,

Page 18: CLARKIN ET AL.BORDERLINE PERSONALITY DISORDER · the personality disorders institute/borderline personality disorder research foundation randomized control trial for borderline personality

BORDERLINE PERSONALITY DISORDER 69

R. C., & Davis, K. L. (1989). Sero-tonergic studies in patients with affec-tive and personality disorders:Correlates with suicidal impulsive ag-gressive behavior. Archives of GeneralPsychiatry, 46, 587-599.

Cookson, A., Espie, J., & Yates, K. (2001). TheEdinburgh Project: A pilot study for thepsychotherapeutic treatment of bor-derline and other severe personalitydisorders. British Journal of Psycho-therapy, 18, 68-88.

Cowdry, R.W., Pickar, D., & Davies, R. (1985).Symptoms and EEG findings in theborderline syndrome. InternationalJournal of Psychiatry in Medicine, 15,201-211.

Crits-Christoph, P. (1998). Psychosocialtreatments for personality disorders. InP. Nathan & J. Gorman (Eds.), A guideto treatments that work. New York: Ox-ford University Press.

Crits-Christoph, P., Cooper, A., & Luborsky,L. (1988). The accuracy of therapists’interpretation and the outcome of dy-namic psychotherapy. Journal of Con-sulting and Clinical Psychology, 56,490-495.

Depue, R. A., & Lenzenweger, M. F. (2001). Aneurobehavioral dimensional model. InW. J. Livesley (Ed.), Handbook of per-sonality disorders (pp. 136-176). NewYork: Guilford.

Depue, R. A., & Spoont, M. R. (1986). Concep-tualizing a serotonin trait: A behavioraldimension of constraint. Annals of theNew York Academy of Sciences, 487,47-62.

Derryberry, D., & Rothbart, M. K. (1988).Arousal, affective, and attentional com-ponents of temperament. Journal ofPersonality and Social Psychology, 55,958-966.

Elkin, I., Pilkonis, P., Docherty, J. P., &Sotsky, S. M. (1988). Conceptual andmethodological issues in comparativestudies of psychotherapy andpharmacotherapy: I. Active ingredientsand mechanisms of change. AmericanJournal of Psychiatry, 145(8), 909-917.

Evans, K., Tyrer, P., Catalan, J., Schmidt, U.,Davidson, K., Dent, J., Tata, P., Thorn-ton, S., Barber, J., & Thompson, S.(1999). Manual-assisted cognitive-be-haviour therapy (MACT): A randomizedcontrolled trial of a brief interventionwith bibliotherapy in the treatment of

recurrent deliberate self-harm. Psycho-logical Medicine, 29, 19-25.

First, M.B., Gibbon, M., Spitzer, R.L., & Wil-liams, J.B.W. (1996). User’s guide forthe structured clinical interview forDSM-IV axis I—research version(SCID-1, Version 2.0, February 1996,Final Version).

Fonagy, P., Leigh, T., Steele, M., Steele, H.,Kennedy, R., Mattoon, G., Target, M., &Gerber, A. (1996). The relation of at-tachment status, psychiatric classifi-cation and response to psychotherapy.Journal of Consulting & Clinical Psy-chology, 64, 22-31.

Fyer, M. R, Frances, A. J., Sullivan, T., Hurt,S. W., & Clarkin, J. F. (1988). Suicideattempts in patients with borderlinepersonality disorder. American Journalof Psychiatry,145(6), 737-739.

Goldstein, H. (1995). Multilevel statisticalmodels (2nd ed). New York: Wiley.

Grinker, R. R, Werble, B., & Drye, R. C.(1968). The borderline syndrome: A be-havioral study of ego-functions. NewYork: Basic Books.

Gunderson, J. G. (1996). Borderline patient’sintolerance of aloneness: Insecure at-tachments and therapist availability.American Journal of Psychiatry, 153,752-758.

Gunderson, J.G., Berkowitz, C., &Ruiz-Sancho, A. (1997). Families ofborderline patients: Apsychoeducational approach. Bulletinof the Menninger Clinic, 66, 446-457.

Gunderson, J.G., & Wheelis, J. (1999). Drs.Gunderson and Wheelis reply. Ameri-can Journal of Psychiatry, 156, 976.

Gurvits, I.G., Koenigsberg, H.W., & Siever,L.J. (2000). Neurotransmitter dysfunc-tion in patients with borderline person-ality disorder. In J. Paris (Ed.),Borderline personality disorder. ThePsychiatric Clinics of North America,W.B. Saunders, Philadelphia, PA, Vol-ume 23 (1), 27-40.

Guthrie, E., Kapur, N., Mackway-Jones, K.,Chew-Graham, C., Moorey, J., Mendel,E., Marino-Francis, F., Sanderson, S.,Turpin, C., Boddy, G., & Tomenson, B.(2001). Randomised controlled trial ofbrief psychological internvention afterdeliberate self poisoning. British Medi-cal Journal, 323, 1-4.

Higgitt, A. & Fonagy, P. (1992). Psychother-apy in borderline narcissistic personal-

Page 19: CLARKIN ET AL.BORDERLINE PERSONALITY DISORDER · the personality disorders institute/borderline personality disorder research foundation randomized control trial for borderline personality

70 CLARKIN ET AL.

ity disorder. British Journal ofPsychiatry, 161, 23-43.

Kass, F., Skodol, A., Charles, E., Spitzer, R., &Williams, J. (1985). Scaled ratings ofDSM-III personality disorders. Ameri-can Journal of Psychiatry, 142 ,627-630.

Kernberg, O.F. (1984). Severe personality dis-orders: Psychotherapeutic strategies.New Haven: Yale University Press.

Kernberg, O.F. (1996). A psychoanalytic the-ory of personality disorders. In J. F.Clarkin, & M. Lenzenweger (Eds.), Ma-jor theories of personality disorder (pp.106-137). New York: Guilford Press.

Kochanska, G. (1995). Children’s tempera-ment, mothers’ discipline, and securityof attachment: Multiple pathways toemerging internalization. Child Devel-opment, 66, 597-615.

Kochanska, G., Murray, K., & Coy, K. C.(1997). Inhibitory control as a contrib-utor to conscience in childhood: Fromtoddler to early school age. Child Devel-opment, 68, 263-277.

Koons, C. R., Robins, C. J., Tweed, J. L.,Lynch, T. R., Gonzalez, A. M., Morse, J.Q., et al. (2001). Efficacy of dialecticalbehavior therapy in women veteranswith borderline personality disorder.Behavior Therapy, 32, 371-390.

Lenzenweger, M.F., Johnson, M.D., & Willett,J.B. Individual growth curve analysisilluminates stability and change in per-sonality disorder features: The longitu-dinal study of personality disorders.(Manuscript submitted for publication,Archives of General Psychiatry).

Lenzenweger, M. F., Loranger, A. W., Korfine,L., & Neff, C. (1997). Detecting person-ality disorders in a nonclinical popula-tion: Application of a 2-stage caseidentification. Archives of General Psy-chiatry, 54(4), 345-351.

Levy, K.N., Clarkin, J.F., Foelsch, P.A., &Kernberg, O.F. (2003). Transfer-ence-focused psychotherapy for pa-tients diagnosed with borderlinepersonality disorder: A comparisonwith a treatment-as-usual cohort.Manuscript submitted for publication.

Liberman, R. P., & Eckman, T. (1981). Behav-ior therapy vs. insight-oriented therapyfor repeated suicide attempters. Ar-chives of General Psychiatry, 38,1126-1130.

Linehan, M. M. (1993). Cognitive-behavioraltreatment of borderline personality dis-order. New York: Guilford Press.

Linehan, M. M, Armstrong, H. E., Suarez, A.,Allmon, D., & Heard, H. (1991). Cogni-tive-behavioral treatment of chroni-cally parasuicidal borderline patients.Archives of General Psychiatry,48,1060-1064.

Linehan, M.M.,Heard, H.L., & Armstrong,H.E. (1993). Naturalistic follow up of abehavioral treatment for chronicallyparasuicidal borderline patients. Ar-chives of General Psychiatry, 50,971-974.

Linehan, M. M., Schmidt, H. I., Dimeff, L. A.,Craft, J. C., Kanter, J., & Comtois, K. A.(1999). Dialectical behavior therapy forpatients with borderline personalitydisorder and drug-dependence. Ameri-can Journal on Addictions, 8, 279-292.

Links, P. S., Heslegrave, R., & van Reekum, R.(1999). Impulsivity: Core aspect of bor-derline personality disorder. Journal ofPersonality Disorders, 13, 1-9.

Loranger, A.W. (1999). International Personal-ity Disorder Examination (IPDE) man-ual. Odessa, FL: Psychological Assess-ment Resources, Inc.

Luborsky, L., Diguer, L., Seligman, D.,Rosenthal, R., Krause, E., Johnson, S.,Halperin, G., Bishop, M., Berman, J.S.,& Schweizer, E. (1999). The re-searcher’s own therapy allegiances: A“wild card” in comparisons of treatmentefficacy. Clinical Psychology: Scienceand Practice, 6, 95-106.

McGlashan T. H. (1986). The Chestnut Lodgefollow up study: III. Long-term outcomeof borderline personalities. Archives ofGeneral Psychiatry, 43, 20-30.

Mandell, A., Knapp, S., Ehlers, C., & Russo,P.V. (1984). The stability of constrainedrandomness: Lithium prophylaxis atseveral neurobiological levels. In R.Post & J. C. Ballenger (Eds.),Neurobiology of mood disorders (pp.774-776). Baltimore, MD: Williams &Williams.

Mann, J. J., Waternaux, C., Haas, G. L., &Malone, K. M. (1999). Toward a clinicalmodel of suicidal behavior in psychiat-ric patients. American Journal of Psy-chiatry, 156, 181-189.

Masterson, J.R. (1981). The narcissistic andborderline disorders: An integrated de-velopmental approach. New York:Brunner/Mazel.

Page 20: CLARKIN ET AL.BORDERLINE PERSONALITY DISORDER · the personality disorders institute/borderline personality disorder research foundation randomized control trial for borderline personality

BORDERLINE PERSONALITY DISORDER 71

Moeller, F.G., Barratt, E.S., Dougherty, D.M.,Schmitz, J.M., & Swann, A.C. (2001).Psychiatric aspects of impulsivity.American Journal of Psychiatry, 158,1783-1793.

Munroe-Blum, H., & Marziali, E. (1995). Acontrolled trial of short-term grouptreatment for borderline personalitydisorder. Journal of Personality Disor-ders, 9, 190-198.

Nurnberg, H. G., Raskin, M., Levine, P. E.,Pollack, S., Siegel, O., & Prince, R.(1991). The comorbidity of borderlinepersonality disorder and otherDSM-III-R Axis II personality. AmericanJournal of Psychiatry, 148, 1371-1377.

Oldham, J. M., Gabbard, G. O., Goin, M. K.,Gunderson, J., Soloff, P., Spiegel, D.,Stone, M., & Phillips, K. A. (2001).Practice guideline for the treatment ofpatients with borderline disorder.American Journal of Psychiatry [suppl],158,1-52.

Perry, J. C., Banon, E., & Ianni, F. (1999). Ef-fectiveness of psychotherapy for per-sonality disorders. American Journal ofPsychiatry, 156, 1312-1321.

Piper, W. E., Joyce, A. S., McCallum, M., &Azim, H. F. (1998). Interpretive andsupportive forms of psychotherapy andpatient personality variables. Journalof Consulting & Clinical Psychology, 66,558-567.

Posner, M. I., & Rothbart, M. K. (1988). Atten-tion, self-regulation and conscious-ness. Philosophical Transactions of theRoyal Society of London-Series B: Bio-logical Sciences, 353 (1377), 1915-27.

Posner, M. I., & Rothbart, M. K. (2000). Devel-oping mechanisms of self-regulation.Developmental Psychopathology, 12,427-441.

Posner, M. I., Rothbart, M. K., Vizueta, N.,Levy, K., Thomas, K. M., & Clarkin, J.(2002). Attentional mechanisms of bor-derline personality disorder. Proceed-ings of the National Academy ofSciences, USA, 99, 16366-16370.

Posner, M.I., Rothbart, M., Vizueta, N.,Thomas, K., Levy, K., Fossella, J.,Silbersweig, D., Stern, E., Clarkin, J., &Kernberg, O. (2003). An approach tothe psychobiology of personality disor-ders. Development andPsychopathology, 15, 1093-1106.

Raudenbush, S., & Bryk, A. (2001). Hierarchi-cal linear models (2nd ed). ThousandOaks, CA: Sage Publications.

Rockland, L.H. (1992). Supportive therapy forborderline patients. New York: Guilford.

Rosenthal, R., & Rosnow, R.L. (1991). Essen-tials of behavioral research: Methodsand data analysis (2nd ed). New York:McGraw-Hill.

Rosenthal, R., Rosnow, R.L., & Rubin, D.B.(2000). Contrasts and effect sizes in be-havioral research: A correlational ap-proach. New York: CambridgeUniversity Press.

Rothbart, M. K., Ahadi, S. A., & Hershey, K. L.(1994). Temperament and social be-havior in childhood. Merrill-PalmerQuarterly, 40, 21-39.

Ryle, A., & Golynkina, K. (2000). Effectivenessof time-limited cognitive analytic ther-apy of borderline personality disorder:Factors associated with outcome. Brit-ish Journal of Medical Psychology, 73,197-210.

Salkovskis, P. M., Atha, C., & Storer, D.(1990). Cognitive-behavioural problemsolving in the treatment of patients whorepeatedly attempt suicide: A con-trolled trial. British Journal of Psychia-try, 157, 871-876.

Scheel, K. R. (2000). The empirical basis of di-alectical behavior therapy: Summary,critique, and implications. Clinical-Psy-chology: Science-and-Practice,7(1),68-86. Cary, NC: Oxford UniversityPress.

Siever, L. J., & Davis, K. L. (1991). Apsychobiological perspective on thepersonality disorders. American Jour-nal of Psychiatry, 148(12), 1647-1658.

Siever, L. J., & Trestman, R. L. (1993). The se-rotonin system and aggressive person-ality disorders. International ClinicalPsychopharmacology, 8 (suppl.),2:33-39.

Silbersweig, D. A., Pan, H., Beutel, M., Ep-stein, J., Goldstein, M., Thomas, K.,Posner, M., Hochberg, H., Brendel, G.,Yang, Y., Kernberg, O., Clarkin J., &Stern, E. (2001, January).Neuroimaging of inhibitory and emo-tional function in borderline personal-ity disorder. Presented at the AmericanCollege of Neuropsychopharmacology.

Silbersweig, D. A., & Stern, E. (2001). Func-tional neuroimaging andneuropsychology: Convergence, ad-vances and new directions. Journal ofClinical Experimental Neuropsychology,23,1-2.

Page 21: CLARKIN ET AL.BORDERLINE PERSONALITY DISORDER · the personality disorders institute/borderline personality disorder research foundation randomized control trial for borderline personality

72 CLARKIN ET AL.

Silverman, J. M., Pinkham, L., Horvath, T. B.,Coccaro, E. F., Klar, H., Schear, S.,Apter, S., Davidson, M., Mohrs, R. C., &Siever, L. J. (1991). Affective and im-pulsive personality disorder traits inthe relatives of patients with borderlinepersonality disorder. American Journalof Psychiatry, 148, 1378-1385.

Singer, J.D., & Willett, J.B. (2003). Appliedlongitudinal data analysis: Modelingchange and event occurrence. NewYork: Oxford University Press.

Skodol, A., Gunderson, J., Pfohl, B., Widiger,T. A., Livesley, W. J., & Siever, L. J.(2002). The borderline diagnosis II: Bi-ology, genetics, and clinical course. Bi-ological Psychiatry, 51, 951-963.

Soloff, P. H. (2000). Psychopharmacology ofborderline personality disorder.Psychiatria Clinica North America,23,169-192.

Soloff, P.H., Lynch, K.G., & Kelly, T. M. (2002).Childhood abuse as a risk factor forsuicidal behavior in borderline person-ality disorder. Journal of PersonalityDisorders, 16 (3), 201-214.

Spoont, M. R. (1996). Emotional instability.In C. G. Costello (Ed.), Personality char-acteristics of the personality disordered(pp. 48-90). New York: Wiley.

Spoont, M. R. (1992). Modulatory role of sero-tonin in neutral information process-ing: Implications for humanpsychopathology. Psychological Bulle-tin, 112, 330-350.

Steiner, J. (1993). Psychic retreats. New York:Routledge.

Stevenson, J., & Meares, R. (1992). An out-come study of psychotherapy for pa-tients with borderline personalitydisorder. American Journal of Psychia-try, 149(3), 358-362.

Stone, M. H. (1993). Long-term outcome inpersonality disorders. British Journal ofPsychiatry, 162, 299-313.

Swenson, C. (1989). Kernberg and Linehan:Two approaches to the borderline pa-tient. Journal of Personality Disorder,31, 26-35.

Torgersen, S. (1984). Genetic and nosologicalaspects of schizotypal and borderlinepersonality disorders: A twin study. Ar-chives of General Psychiatry, 41,546-554.

Torgersen, S., Kringlen, E., & Cramer, V.(2001). The prevalence of personalitydisorders in a community sample. Ar-chives of General Psychiatry, 58,590-596.

Torgersen, S., Lygren, S., Oien, P. A., Skre, I.,Onstad, S., Edvardsen, J., Tambs, K.,& Kringlen, E. (2000). A twin study ofpersonality disorders. ComprehensivePsychiatry, 41, 416-425.

Waldinger, R.J., & Gunderson, J.G. (1989).Completed psychotherapies with bor-derline patients. American Journal ofPsychotherapy, 38, 190-201.

Wampold, B. E. (2001). The great psychother-apy debate: Models, methods, and find-ings. Mahwah, NJ: Lawrence Erlbaum.

Westen, D. (2000). The efficacy of dialecticalbehavior therapy for borderline person-ality disorder. Clinical Psychology: Sci-ence and Practice, 7, 92-94.

Widiger, T. A, & Weissman, M. M. (1991). Epi-demiology of borderline personality dis-order. Hospital Community Psychiatry,42, 1015-1021.

Zanarini, M.C., Frankenburg, F., Hennen, J.,& Silk, K. (2003). The longitudinalcourse of borderline pathology: 6-yearprospective follow up of the phenomen-ology of borderline personality disor-der. American Journal of Psychiatry,160, 274-283.

Zanarini, M. C. (1993). BPD as an impulsespectrum disorder. In J. Paris (Ed.),Borderline personality disorder: Etiol-ogy and treatment (pp.67-85). Wash-ington, DC: American PsychiatricPress.

Zimmerman, M., & Coryell, W. (1990). Diag-nosing personality disorders in thecommunity. A comparison of self-re-port and interview measures. Archivesof General Psychiatry, 47,527-531.